Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-04-01 Cooperative journals: 《计算机应用研究》
Abstract： Traditional content-based recommendation algorithms have lower accuracy, while data sparseness and cold start problems are common in collaborative filtering recommendation algorithms. To solve this problem, this paper proposed a hybrid recommendation algorithm based on content and collaborative matrix factorization technique. The algorithm realized the decomposition of content and collaborative matrix in a common low-dimensional space while preserving the local data structure. This paper used an iterative method based on multiplication update rules in parameter optimization, improved learning ability. The experimental results show that the proposed algorithm is superior to other representative projects cold start recommendation algorithm, which effectively alleviates the data sparseness and improves the efficiency of the algorithm.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-01-28 Cooperative journals: 《计算机应用研究》
Abstract： To overcome the local optimum and premature convergence due to loss of population diversity of the particle swarm optimization algorithm, this paper proposed a disruption particle swarm optimization algorithm based on exponential decay weight (EDW-DPSO) . Firstly, the population was semi-uniformly initialized to distribute the population in an overall uniform, locally random manner. Secondly, the dynamic splitting operator was introduced to perform splitting operations on particles which satisfying the splitting condition, increasing the diversity of the population and avoiding the particles falling into local optimum. Finally, the exponential decreasing inertia weight was used to balance the global search and local development ability of the particles. The experimental results show that the algorithm has a large search space in the early stage, and the population diversity increases. In the later stage, emphasizeing the local development to improve the convergence precision and optimization ability. It can also accelerate particles jumping out of the local extremum and approximate globle optimum.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-10-11 Cooperative journals: 《计算机应用研究》
Abstract： Association rules are attracting wide attention in the field of big data mining. The key and difficult point of the algorithm is to mine frequent sets. In order to further improve the speed of the association rules mining frequent sets and optimize the execution performance of the algorithm, an association rule mining algorithm based on improved memory structure is proposed. For the existing algorithm, the storage structure is simple, the candidate set with a large amount of redundancy is generated, the time and space complexity is high, and the mining efficiency is not ideal. The algorithm of this paper is based on the Spark distributed framework. The partitions are mined in parallel to extract frequent sets. It is proposed to use the Bloom filter to store the project in the mining process, and to simplify the operation of the transaction set and the candidate set, so as to optimize the speed of mining frequent sets. Save computing resources. Compared with the YAFIM algorithm and the MRApriori algorithm, the algorithm has a significant improvement in the efficiency of mining frequent sets under the condition of occupying less memory. The algorithm can not only improve the mining speed, reduce the memory pressure, but also has good scalability, so that the algorithm can be applied to larger data sets and clusters, so as to optimize the performance of the algorithm.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-07-09 Cooperative journals: 《计算机应用研究》
Abstract： In order to solve the problems such as sparse data, cold start and lack of diversity of recommendation results in traditional collaborative filtering recommendation algorithms, this paper proposes a collaborative filtering recommendation algorithm that integrates social networking with key users. Based on the score matrix of user projects, the algorithm integrates user social networks to derive social trust matrix, integrates key user information to obtain key user scoring matrix, and then uses these three matrix data distributions to predict user's target project with different weights. score. At the same time, aiming at the massive data problem, this paper uses the Spark distributed cluster to realize the parallelization of the algorithm. The experimental results show that the algorithm can effectively alleviate the data sparse problem and improve the data processing speed and recommendation accuracy.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-04-12 Cooperative journals: 《计算机应用研究》
Abstract： The Word2Vec model is a neural network model (NNLM) that converts words in text into a word vector. It is widely used in natural language processing tasks such as emotional analysis, question answering robot and so on. Word vectors generated for the Word2Vec model lacked the ambiguity of context and the inability to create OOV word vectors. Based on the similarity information of document context and Word2Vec model, this paper proposed a word vector generation model that conforms to the meaning of OOV context. It is called the Word2Vec-ACV model. The model was similar to the process of the word vector generated by the Word2Vec model, but it was different. First of all, Word2Vec model of the continuous word bag (CBOW) and the Hierarchical Softmax trained the word vector matrix, namely the weight matrix. Secondly, the co-occurrence matrix was normalized to get the average context word vector. Then, the word vector consisted of an average context word vector matrix. Finally, the vector matrix of the average context word vector matrix and the weight matrix were multiplied to get the word vector matrix. In order to simultaneously solved the ambiguity problem of out of vocabulary words and out of vocabulary words to create. In this paper, the average context word vectors were divided into two kinds: the global average context word vector (global ACV) and the local average context word vector (local ACV) . In addition, the two taken the weight value to form a new average context word vector matrix. The Word2Vec model can effectively express the word in vector form. Experiments on analogical tasks and named entity recognition (NER) tasks respectively, the results show that the Word2Vec-ACV model is superior to the Word2Vec model in the accurate expression of the word vector. It is a word vector representation method to create a contextual context for OOV words.
Hits
Hits
Downloads
Comment